SELECTIVE RETINA THERAPY (SRT): A REVIEW ON

METHODS, TECHNIQUES, PRECLINICAL AND FIRST

CLINICAL RESULTS

BRINKMANN R.1, ROIDER J.2, BIRNGRUBER R.1

ABSTRACT

Selective retina therapy (SRT) is a new laser procedure for retinal diseases that are thought to be associ-ated with a degradation of the retinal pigment epithelium (RPE). The aim of the irradiation is to selectivelydamage the RPE without affecting the neural retina, the photoreceptors and the choroid. Goal of the treat-ment is to stimulate RPE cell migration and proliferation into the irradiated areas in order to improve themetabolism at the diseased retinal sites. In a pilot study more than 150 patients with soft drusen, retino-pathia centralis serosa (RCS) and macular edema were treated. The first 3-center international trial targetsdiabetic macular edema and branch vein occlusion.In this review, selective RPE effects are motivated and two modalities to achieve selective RPE effects willbe introduced: a pulsed and a continuous wave scanning mode. The mechanism behind selective RPE-ef-fects will be discussed reviewing in vitro results and temperature calculations. So far clinical SRT is per-formed by applying trains of 30 laser pulses from a Nd:YLF-Laser (527 nm, 1.7 µs, 100 Hz) to the dis-eased fundus areas. In the range of 450-800 mJ/cm2 per pulse, RPE-defects in patients were provedangiographically by fluorescein or ICG-leakage. The selectivity with respect to surrounding highly sensitivetissue and the safety range of the treatment will be reviewed. With the laser parameters used neither bleed-ing nor scotoma, proved by microperimetry, were observed thus demonstrating no adverse effects to thechoroid and the photoreceptors, respectively.During and after irradiation, it shows that the irradiated locations are ophthalmoscopically invisible, sincethe effects are very limited and confined to the RPE, thus a dosimetry control is demanded. We report ona non-invasive opto-acoustic on-line technique to monitor successful RPE-irradiation and compare the datato those achieved with standard angiography one-hour post treatment.

KEYWORDS

Selective treatment, RPE, µs-pulses, on-line dosimetry, laser scanner, AMD

RÉSUMÉ

La thérapie rétinienne sélective (TRS) est un nouveau procédé laser pour les maladies rétiniennes que l’onsuppose être liées à une dégradation de l’épithélium pigmentaire rétinien (EPR). Le but de l’irradiation estd’endommager sélectivement l’EPR sans affecter la rétine neurale, les photorécepteurs et la choroïde.L’objectif du traitement est de stimuler la migration et la prolifération des cellules de l’EPR dans les zonesirradiées en vue d’améliorer le métabolisme au niveau des sites rétiniens affectés. Dans une étude pilote,plus de 150 patients atteints d’un épaississement diffus, de rétinopathie séreuse centrale (RSC) et d’œdèmemaculaire ont été traités. Le premier essai international tricentrique vise l’œdème maculaire diabétique etl’occlusion des branches veineuses.

zzzzzz

1 Medical Laser Center Lübeck, Germany2 University Eye Clinic Kiel, Germany

51Bull. Soc. belge Ophtalmol., 302, 51-69, 2006.

Dans ce compte-rendu, des effets sélectifs de l’EPR sont mentionnés et deux procédures permettant d’obtenirdes effets sélectifs de l’EPR seront introduits: un mode de balayage à ondes pulsées et un mode de balay-age à ondes continues. Le mécanisme provoquant les effets sélectifs de l’EPR sera décrit en examinant lesrésultats in vitro et les calculs de température. Ainsi une TRS clinique est menée en appliquant des trainsde 30 impulsions laser à partir d’un laser Nd:YLF (527 nm, 1.7 µs, 100 Hz) au niveau des zones affectées.Dans la gamme de 450 à 800 mJ/cm2 par impulsion, des anomalies de l’EPR ont été constatées de manièreangiographique par une fuite de la fluorescéine ou du vert d’indocyanine. La sélectivité par rapport auxtissus adjacents hautement sensibles et l’intervalle de sécurité du traitement seront examinés. Avec lesparamètres laser utilisés, ni saignement ni scotome, ceci prouvé par micropérimétrie, n’ont été observés,ce qui indique ainsi l’absence d’effets indésirables sur la choroïde et les photorécepteurs, respectivement.Pendant et après l’irradiation, il est démontré que les locations irradiées sont invisibles d’un point de vueophtalmoscopique, les effets étant très limités et confinés à l’EPR, ainsi un contrôle de la dosimétrie estrequis. Nous décrivons une technique optoacoustique non invasive en ligne permettant de contrôler l’irradiationréussie de l’EPR et de comparer les données avec celles obtenues avec une angiographie standard uneheure après le traitement.

MOTS-CLÉS

Traitement sélectif, EPR, impulsions µs, dosimétrie en ligne, scanner laser, DMLA

52

INTRODUCTION

Laser photocoagulation of the retina has been performed for more than 30 years and is wellestablished for various retinal diseases. Most commonly used are laser wavelengths in the greenspectral range (Ar+-Laser: 514 nm, Nd-Laser: 532 nm), however, also red and near IR-wave-lengths are customary. Irradiation is performed typically with a power of 50-300 mW appliedover a time 50-300 ms per treated area/spot. The main absorber of the light at the fundus is theretinal pigment epithelium (RPE), which is heavily loaded with strong light absorbing melano-somes (Fig. 1). With the irradiation parameters irreversible thermal denaturation is induced tothe RPE 3,56 and owing to heat flow from the RPE into the surroundings also to the choroid andthe inner and outer segments of the retina.56 The power and time of the laser irradiation is ad-justed by the ophthalmologist to produce a gray or whitish retinal lesion, originating from in-creased light scattering indicating tissue denaturation at the irradiation site. The grayish lesionsare also regarded as a visible endpoint of a successful photocoagulation and thus serve as adosimetry control.Following laser photocoagulation, the targeted tissue undergoes a healing process. Typically, thetissue in the target area will be replaced by proliferating glial tissue originating from the sur-rounding retina and choroid. RPE cells also contribute significantly to this healing process.58 Ithas been shown in animal experiments that the RPE responds in several ways after injury. RPEcells adjacent to the irradiation site may spread out and cover the defect by hypertrophy of neigh-boring RPE cells.9,43 The division of RPE cells has also been shown after photocoagulation inrabbits,24 in monkeys after retinal detachment,24 and in rabbits after surgically induced RPEdefects.23 The outer blood retinal barrier is normally restored after about 7 days.25 Further, Gla-ser has shown that RPE cells produce inhibitors for neovascularization, suggesting that thesecells may play a role in the regulation of new vessel growth.20 In addition, Boulton et al. foundsignificant change of growth factors in the vitreous after panretinal photocoagulation.4 Yoshimu-ra et al. showed that photocoagulated RPE cells produce inhibitors of endothelial cells.61 Themolecular and immunologic characteristics of these inhibitors correlate with the TGF-β2. Regen-erating RPE cells are known to produce more TGF-β2 compared to normal RPE cells.34

The benefit of retinal laser photocoagulation in various diseases has traditionally been attributedto the destruction of retinal tissue. However, the exact biological effect and the physiological

Fig. 1. Transmission electron microscopy (TEM) of a human RPE cell in its environment.18

53

improvement leading to the desired therapeutic effect after this locally severe retinal damage ispoorly understood.In the treatment of diabetic macular edema, the beneficial effect is thought to be mediated bythe restoration of a new RPE barrier.5 A similar effect can be postulated in the treatment of drus-en which are located within Bruch’s membrane or beneath the RPE and often disappear afterphotocoagulation of the nearby tissues, eventually likely as a byproduct of surrounding phago-cytosis. The efficacy of drusen treatment was studied by several investigators.14,16 However, sincethe drusen are located below the retina, there is no clear rationale to routinely include the neuralretina in the photocoagulation. The same arguments may count generally in the treatment ofmacular edema. For instance, in central serous retinopathy (CSR) the rationale of therapy is thephotocoagulation and subsequent formation of a new metabolically improved RPE barrier. Here,the simultaneous destruction of the photoreceptors can be regarded as an unwanted side effect.The most common explanation for the beneficial effect of photocoagulation in diabetic retinop-athy is the destruction of oxygen consuming photoreceptors.59 However, another theory sug-gests that the beneficial effect results from the restoration of a new RPE barrier. The subsequentproduction of a variety of growth factors 33,60-61 results in an improved RPE metabolism andimproved pump function to resolve the edema.

CONCEPT OF SELECTIVE RETINA TREATMENT

(SRT)

For a variety of retinal diseases, especially those which are thought to be associated with a deg-radation of the RPE, it might be sufficient to selectively damage the RPE while the adjacentphotoreceptors, the neural retina and the choroid can be spared and thus scotoma be avoided.44

This is especially useful and demanded within the macular area. If the damaged RPE is rejuve-nated in the healing process due to migration and proliferation of the adjoining RPE, such aminimal destructive selective RPE treatment might be optimal for therapy.The question arises if and how selective RPE effects can be performed. The concept of selectivetargeting of naturally or artificially pigmented cells or organelles in less strong absorbing sur-roundings was introduced by Anderson and Parrish 1,38 and led to a variety of applications inophthalmology 29,43 and dermatology 38 using pulsed laser radiation. Selective cell effects aremost interesting in cases where highly sensitive cells, which have to be preserved, are close tothe target cells. In this case, a basis for selective RPE damage can be found in the strong lightabsorbing melanosomes inside the RPE cells, which absorb about 50 % of the incident light inthe green spectral range 19 and thus are the dominant chromophores within the fundus of theeye. If laser pulses are applied with pulse durations shorter then the time needed for the pro-duced heat to diffuse, locally confined high temperatures can be obtained at the absorber site.The so-called thermal relaxation time τR can be estimated for a spherical particle of radius r andthe thermal diffusivity κ to τR = r2/4κ. For a melanosome of a radius r = 0.5 µm, the thermalrelaxation time τR(mel) is calculated to τR(mel) ≈ 420 ns using the thermal diffusivity of water,κ = 1.5 x 105 µm2/s. In order to confine high temperatures to the whole RPE cell with a thick-ness of approximately 5 µm, τR(RPE) ≈ 10 µs is estimated.In order to verify such rough estimations for an RPE cell containing multiple discrete absorbers,mathematical models have been developed allowing a more detailed analysis of the spatial andtemporal temperature distribution at the fundus of the eye.2,22,40-41,44 Figure 2 gives the calcu-lated temperature profile inside the RPE and in 5 µm distance at the photoreceptor level. It showsthe temporal temperature profile as calculated after a laser exposure with an argon laser, a rep-etition rate of 500 Hz with single pulse durations of 5 µs.44 It is conceivable that most of theheat is now concentrated within the RPE at the end of the laser pulse. High temperature peaksonly occur within the RPE, highest temperatures are found at and around the single melaningranules. The peak temperature increases proportional with the pulse energy. Due to heat dif-

54

fusion from volumetrically small particles, the peaks are almost cancelled out at just 5 µm dis-tance.44 On the other hand, the average or background temperature increase at the retina isalmost as high as in the RPE. Generally the background temperature depends on the pulse en-ergy and repetition rate, thus the average laser power, and on the overall dimensions of the irra-diated area. The ratio of peak to average temperature can be maximized if repetitive short laserpulses are applied in such a manner that the following laser pulse is applied only after the retinaltissue has had sufficient time to completely cool down to baseline. If the peak temperatureachieved by one or multiple pulses is high enough to cause damage inside the RPE and the aver-age temperature is low enough to avoid retinal photocoagulation, selective RPE effects shouldbe possible.

APPLICATION MODES FOR SRT

The easiest approach to achieve high peak temperatures is to apply appropriate laser pulses viaa commonly available laser slit lamp to the retina. However, if the selective effects in the RPEonly rely on these local high temperatures within the RPE cells,6 this goal should also be obtain-able when rapidly scanning a continuous wave (cw) laser beam across the fundus. An arbitraryirradiation pattern can be either generated with one pulse (or repetitive pulses) of a certain pulseduration or it can be obtained by scanning a small laser spot once (or repetitively) across thesame area with such a speed, that each point is illuminated for the same time as in the pulsedmode.In figure 3, the two possible application modalities are sketched exemplary for circular spotsassuming an illumination time of 1.7 µs. A pulse duration of 1.7 µs in the pulsed mode corre-sponds to an application with a scanned 18 µm spot using a speed of 10.6 m/s. For this illumi-nation time, a temperature increase of 100 °C at the surface of a RPE-melanosome requires aradiant exposure of about 310 mJ/cm2. This corresponds to a pulse energy of 62 µJ for a 160 µmspot or a laser power of 460 mW for an 18 µm spot scanned across the fundus, respectively.7

Advantages and disadvantages as well as details for both SRT techniques and application mo-dalities (slit lamp vs. fundus camera) are extensively discussed in the literature.8 In general, thehigher flexibility of the scanning approach to easily generate arbitrary pattern geometries com-petes with its more complicated set-up and application.

THRESHOLD IRRADIATION FOR RPE CELL DAMAGE

In order to investigate thresholds of RPE damage in the range of pulse durations within the ther-mal relaxation time of the RPE, two different laser systems in the ns to µs time regimen wereused for the pulsed mode: a Q-switched, pulse-stretched Nd:YLF laser with adjustable pulseduration between 250 ns and 3 µs at a wavelength of 527 nm and a Q-switched, frequency

Fig. 2. Transient temperature profile within the RPE and at the photoreceptors (5 µm distance) when applying a train ofµs-laser pulses with a repetition rate of 500 Hz, after Roider.44

55

doubled Nd:YAG laser at a wavelength of 532 nm and 8 ns exposure time.6 For the scanningmode, the beam of an Ar+-laser was scanned across the probes.8 Freshly harvested porcine RPEsamples served as model. RPE cell vitality prior and after irradiation was probed by the fluores-cent dye marker Calcein-AM (Molecular Probes Inc.).6

Figure 4 shows the equivalent dose for 50 % cell damage (ED50) radiant exposure for differentrepetitive exposures to the RPE in the pulsed and scanned mode. When the RPE is irradiatedwith a train of µs-pulses at a pulse repetition rate of 500 Hz, a decrease of the threshold radiantexposure with the number of pulses applied is observed. With a 40 % drop at 100 pulses, thedecrease is most pronounced at a pulse width of 3 µs. The threshold is mostly reduced duringthe first 500 pulses. Using longer pulse series, the damage threshold almost remains constant.Applying 10 scanned exposures, a threshold power of Pth = 569 mW was found, which resultsin an ED50 radiant exposure of Hth = 297 mJ/cm2.8 The threshold strongly decreases toHth = 131 mJ/cm2 when applying 500 scans, while a higher number of scans does not lead toa further threshold decrease in accordance with the pulsed application mode. It shows that thethreshold radiant exposure is generally higher in the scanning mode, depending on the numberof repetitive exposure. Possible reasons for these slight deviations are extensively discussed inthe literature.8 Anyway, both modes show a saturation with higher number of exposures.

ORIGIN OF CELL DAMAGE

The origin of RPE cell damage for repetitive µs laser pulse exposure has extensively been inves-tigated and shall briefly be reviewed. The idea of selective RPE damage has been published byRoider assuming a pure thermal cell damage.44 It refers to cell and protein denaturation andthus cell necrosis due to the high temperature peaks within the RPE-cell as shown in figure 2.According to the thermal damage model of Arrhenius 2 the rate of damage increases linearlywith time and exponential with temperature and is additive for repetitive temperature increase,thus repetitive exposure should be favorable. Further, according to a thermal model, µs-pulsedurations in the range of the thermal relaxation time allow to obtain a certain temperature in-crease for a long time while still having strong temperature decrease towards the retina. How-ever, with respect to the experimental results, the saturation of the threshold radiant exposure asshown in figure 4 is in contradiction to a pure thermal damage, which predicts a continuous

Fig. 3. Possible application modes to achieve µs-laser exposure:7 either applying laser pulses to a large area with acertain pulse duration (e.g. 1.7 µs) or scanning a continuous wave laser beam across the target area (e.g. 18 µm indiameter with a speed of 10.6 m/s, which gives the same exposure time to each point on the target).

56

decrease towards higher pulse numbers. Further, a strong dependence of the threshold radiantexposure on the pulse duration below or close to thermal confinement does only exist at themelanosomes directly. Already at small distances apart, the temperature profile does mostly re-flect the energy absorbed but less the pulse duration. Thus, at least for the immediate cell dam-age observed here or in angiography, other mechanisms have to be taken into account. RPE celldamage due to stress waves was investigated in vitro by Douki et al. They found that cell dam-age is rather dependent on the stress rise time, as on the peak stress, and requires stress tran-sients around 70 bar/ns.10 However, these pressure gradients can only be achieved with pulsedurations close to acoustic confinement conditions, which is below 1 ns for melanosomes.Apart from these effects, a thermo-mechanical damage of the cells has to be taken into account:Lin and Kelly heated micro absorbers in suspension with ps and ns laser pulses and demon-strated vaporization around the particles.31 A threshold radiant exposure for micro bubble for-mation around bovine melanosomes was found to be 55 mJ/cm2 at a wavelength of 532 nm.Irradiating RPE cells with the same radiant exposure, non-viable cells were only found, whenintracellular bubble formation occurred.26 It was concluded that cell death is caused by thermo-mechanical disruption of the cell structure due to the significantly increased cell volume duringbubble lifetime. However, in the highly sensitive retina, strong photo disruptive effects have tobe avoided in order to absolutely prevent choroidal disruption, which might lead to bleeding.We investigated micro bubble formation around isolated porcine melanosomes by fast flash pho-tography as shown in figure 5 and found threshold radiant exposures for bubble formation andcell death to be very close.6,35 Further both thresholds increase with pulse durations up to 3 µs.Experiments and calculations showed that the temperature for the onset of vaporization at thesurface of isolated melanosomes is around 140 °C.6 This vaporization temperature is higher thenthe boiling temperature of water under normal conditions, since the formation of micro bubbles

Fig. 4. RPE ED50 radiant exposures as a function of number of exposures for the scanned mode (upper curve) 8 and thepulsed mode with pulse durations of 0.25, 1 and 3 µs according to Brinkmann,6 respectively.

57

at or around the melanosomes requires to overcome the surface tension of the bubble, whichcorresponds to a pressure of approximately 3 bars for a diameter of 1 µm.6,27

In another study, we investigated RPE cell damage with a vitality stain up to pulse durations of3 ms while simultaneously observing microvaporisation.53 It shows that pressure waves gener-ated by micro bubble formation can very well be measured with hydrophones. Analyzing theacoustic transients allows distinguishing between pressure waves induced by micro bubble for-mation and by thermo-elastic expansion. This further allows using this opto-acoustic approachas an on-line dosimetry control for SRT as discussed below.Summarizing, the most probable effect for cell damage up to pulse durations of 50 µs is microbubble induced photo disruption. When laser pulses with appropriate energy are applied to RPEcells, highest temperatures are induced in and around the melanosomes. If the vaporization tem-perature of the intracellular plasma is reached, which takes place first at the surface of the mel-anosomes, micro bubbles begin to form here. Due to the simultaneous growth of a high numberof micro bubbles, the cell volume transiently increases, most likely disrupting the cell structure.Using 6 µs pulses, Roegener et al. observed RPE micro bubbles by reflectometry and demon-strated cell damage occurring only if at least in one pulse out of the whole pulse train bubbleswere induced.42

SUITABLE LASER PARAMETER FOR SRT

Assuming micro bubble formation as the predominant damage mechanism in the ns to µs timeregimen, the question arises for the most suitable pulse duration for SRT. For selective RPE da-mage one has to be also concerned about safety of the highly sensitive photoreceptors and neu-ral retina at the one and the choroid on the other side. Both should neither be harmed by heatnor by photo disruption such as bubble and pressure effects. In order to prove the extension ofthe thermo mechanical effects with pulse energy and duration we observed bubble lifetime andgrowth with an interferometer setup.36 Porcine RPE melanosomes in suspension were irradiatedwith ns and µs pulses of frequency doubled, Q-switched Nd:YAG and Nd:YLF lasers, respec-tively. Simultaneously, bubble lifetime was observed by continuous wave HeNe probe laser light(633 nm) deflected by the bubbles and bubble dynamics with an interferometer setup. Exem-plary bubble extension was recorded by fast flash photography with a N2-Dye laser pulse (< 4ns) and a video microscope.In the case of ns irradiation the bubble size increases with radiant exposure (Fig. 6, on the left).In the case of 1.8 µs pulses, increasing radiant exposure leads to an earlier onset of bubble for-mation relative to the laser pulse, whereas the lifetime of the bubble remains nearly constant(Fig. 6, on the right). Further, bubble oscillation is observed and the number of bubbles duringthe pulse increases. Between two bubbles an off time can be observed.Analyzing the results more closely, a linear relationship between bubble size, observed by flashphotography, and lifetime is obtained for both pulse durations used in this experiment. All data

Fig. 5. Fast flash photographs of a melanosome before irradiation (left) and at the half of the bubble lifetime (right,bubble boundary is marked by white arrows). Irradiation with a 12 ns laser pulse (532 nm) at 369 mJ/cm2 (right) 36.

58

fit the Rayleigh equation, which is known from inertia limited bubble growth,39 within an inter-val of ± 20 %. Bubble lifetimes of 250 and 500 ns correspond to a maximum bubble diameterof 1 and 2 µm, respectively. The mean threshold radiant exposures for bubble nucleation areH12ns = 120 ± 28 mJ/cm2 and H1.8µs = 620 ± 159 mJ/cm2. All data were collected frommany individual melanosomes. For the ns-exposure, the maximum bubble size correlates withthe pulse energy, while for µs-exposure the onset of bubble formation depends on the pulse ener-gy, its dynamics on the pulse duration.For SRT, it is recommended to use µs-laser pulses close to threshold in order to keep the selec-tivity and to avoid the formation of large bubbles and thus the risk of photo disruption of theretina (scotoma) or the choroid (bleeding). Repetitive pulse application is recommended sincethe threshold radiant exposure for cell damage decreases as shown in figure 4, and thus thepulse energy can be reduced.

SELECTIVITY OF TREATMENT

The questions arise, if such selective RPE damage without harming the photoreceptors and thechoroid can be achieved in vivo, and how large the safety regime is with respect to pulse energyuntil adverse tissue effects occur. The selective damage of the RPE in vivo has first been demon-strated by Roider in rabbits by using 10-500 Ar+-laser pulses of 5 µs in duration at a repetitionrate of 500 Hz.43 Fluorescein angiography was accomplished to visualize RPE defects. It showsthat the treated areas close above threshold for fluorescein leakage were ophthalmoscopicallyinvisible directly after treatment as well as in the follow up period. Two weeks after treatmentthe lesions were recovered by RPE. Four weeks post treatment a morphologically completely

Fig. 6. Time plot of bubble induced deflection of probe laser light for pulse durations of 12 ns (left) and 1.8 µs (right)with increasing radiant exposure (top to bottom). The green pulses show the laser pulse. The duration of dips in thetransmitted probe intensity refer to the lifetime of the bubble.36

59

restored RPE barrier showing normal RPE cells was found.43 The selectivity of damage to RPEcells sparing the photoreceptors was demonstrated by histological examinations at different timesafter treatment.Further in vivo experiments were conducted to explore the angiographic and ophthalmoscopicED50 threshold radiant exposures over pulse duration in rabbits in more detail.15 First of all, weexplored the thresholds by means of a commonly used Argon+-laser with anirradiation time of 200 ms. Angiographic and ophthalmoscopic thresholds were obtained at165 W/cm2 and 205 W/cm2, respectively. Thus, a small window to achieve solely RPE damageseemed feasible. However, Roider also found different ophthalmoscopic and angiographic thresh-olds, which differed by a factor of 2 for long pulse irradiation with a 514 nm argon laser (50 ms,100 ms, 500 ms and 1 s), but histological findings nevertheless revealed destruction of the chor-oid and the photoreceptors.43

With respect to µs exposure times, Fig. 7 shows exemplary the probabilities of angiographic andophthalmoscopic damage for a pulse duration of 1.7 µs over the peak radiant exposure in rab-bits. The angiographic threshold decreases towards shorter pulse duration from 189 mJ/cm2 (5µs) to 143 mJ/cm2 (1,7 µs) and 97 mJ/cm2 (200 ns), respectively. A comparable decrease overpulse duration was also found for RPE damage in vitro using a vitality stain to prove cell dam-age, as shown in figure 4.6 The threshold decrease is expected and can be explained with heatdiffusion during irradiation towards longer pulses, since thermal confinement is lost and thushigher pulse energy is needed to reach the vaporization threshold.The described monotonic increase of the angiographic threshold with pulse duration was notobserved for the ophthalmoscopic thresholds. The ED50-thresholds ranged from 478 mJ/cm2

(1,7 µs), 362 mJ/cm2 (5 µs) to 438 mJ/cm2 (200 ns) showing no significant correlation withthe pulse duration. A reason could be a higher inaccuracy to determine the whitening of theretina. It is also known that during the first hour after irradiation a biological enhancement of thelaser effect, e.g. due to occurrence of an intra- or intercellular edema appears.32 This might leadto difficulties to judge an opthalmoscopically visible laser lesion, especially for lesions appliedwith just small energies.For SRT it is important to have an adequate threshold radiant exposure ratio between ophthal-moscopic and angiographic thresholds, so called ’’selective window’’ to prevent surrounding tis-sue damage. Because of the intra- as well as the inter individual pigmentation, which varies bya factor of 2 in healthy humans 19 and most likely more in diseased areas, a factor of at least 2

Fig. 7. Angiographic and ophthalmoscopic damage probability for Nd:YLF-laser exposure (wavelength: 527 nm, 100pulses applied with a repetition rate of 500 Hz) in rabbits with a pulse duration of 1.7 µs on a retinal spot of 102 µmin diameter15. Laser irrdiation was performed in five eyes with completely 137 spots. From the linear fit, the 50 %probabilities are obtained: ED50,opht. = 391 mJ/cm2, ED50,ang. = 113 mJ/cm2.

60

must be ensured for safe clinically treatment without dosimetry control. Damage of Bruch’s mem-brane by short pulses should also be avoided because it is known that new vessel formation canbe promoted.11,30

Referring to the rabbit data we found a decreasing threshold ratio with increasing pulse durationas expected. The ED50 threshold ratio for a pulse duration of 1.7 µs is 3.3.15 However, this ratiois reduced to a factor of about 2 when taking into account the safety ratio between the ED99angiographic and ED1 ophthalmoscopic threshold. For all parameters used we never found rup-tures or hemorrhages below the ED86 ophthalmoscopic threshold. However, histological evalu-ations have to prove whether no extensive mechanical rupture, due to fast bubble extension,lead to a photoreceptor damage. Such damage might not necessarily be ophthalmoscopicallyvisible. The data give a fundamental basis towards appropriate settings for selective RPE treat-ment, however, the threshold data and ratios for selectivity in rabbits can not simply be trans-ferred to the human retina. Thus, careful clinical investigations were needed as discussed in thefollowing paragraph.

SELECTIVITY OF SRT WITH RESPECT TO AVOID

SCOTOMA

In the pilot study, test laser lesions were used to investigate whether sparing of the retina waspossible in the human retina.46 To investigate whether the RPE effects were really selective,microperimetry was performed on 17 patients operating the Nd:YLF laser (527 nm) with a pulseduration of 1.7 µs and a repetition rate of 500 Hz, either applying 100 or 500 pulses per trea-tment spot. Out of 179 test lesions, 73 were followed at various time steps up to one year followup by performing microperimetry directly on top of the laser lesions. For testing, the laser lesionsthreshold stimuli were determined before laser exposure. The threshold sensitivity values weredefined as the minimal contrast at which a response was obtained. This threshold value wasused for evaluating the test lesions in the follow-up period.None of the laser effects were visible by ophthalmoscopy during or directly after SRT, while fluo-rescein angiography clearly demarked the lesions. One day after treatment, retinal defects couldbe detected in up to 73 % of the patients treated with 500 pulses at 100 µJ. Most of thesedefects were no longer detectable after three months. After exposure with 100 pulses no defectscould be found with 70 and 100 µJ after one day and the neural retina remained undamaged inthe follow-up period.46,48

The onset of thermal damage after applying 500 pulses with 500 Hz can be attributed to a strongincreasing background temperature, which does not only depend on the pulse energy, it alsoincreases with higher pulse repetition rate and larger spot diameter due to reduced heat diffu-sion.2 In order to avoid high average temperature increase during irradiation, we irradiated allpatients, apart from the very first ones, with just 100 pulses at a repetition rate of 500 Hz on aretinal spot of 160 µm. In the ongoing multicenter trial we slightly increased the spot diameterto 200 µm, but therefore reduced the pulse repetition rate to 100 Hz. In order to avoid eye mo-vements during treatment, the irradiation time was limited to 300 ms, referring to 30 pulses perspot.The microperimetry findings impressively demonstrate the selectivity after SRT in contrast toconventional laser photocoagulation. Laser photocoagulation is associated with thermal damageof the outer and inner nuclear layer and replacement by scar tissue.45,57-58 It is not surprisingthat such lesions lead to absolute scotoma on microperimetry, which can be detected over eachargon laser exposure. If such lesions are located temporal to the fovea the patient might becomereading problems, despite a good visual acuity.

61

FIRST CLINICAL RESULTS

A salient question is whether and in which diseases selective RPE treatment might lead to apositive therapeutic effect. In a first clinical study we focused on three pathological conditions:diabetic macular edema, central serous retinopathy, and drusen in age related macular degene-ration (AMD). Twelve patients with diabetic maculopathy (group I), ten with soft drusen (groupII) and four with central serous retinopathy (CSR) (group III) were treated and followed up forone year.47

SRT was performed with a self developed frequency doubled, pulse stretched Nd:YLF laser (wa-velength 527 nm, pulse duration 1.7 µs).6 Either 30 laser pulses per area with a repetition rateof 100 Hz or 100 pulses at 500 Hz, respectively, were applied on a spot diameter of 160 µm.Angiographically visible lowest radiant exposures were found between 350-500 mJ/cm2 per pul-se in test expositions at the arcades to prove for individual dosimetry among different patients.For the treatment typically 650 mJ/cm2 were used, in order to compensate for the variation inpigmentation and ocular transmission.Patients were examined at various times after treatment by ophthalmoscopy, fluorescein- andICG angiography as well as infrared imaging. After six months in group I hard exsudates disap-peared in 6 out of 9, and leakage disappeared in 6 out of 12 diabetic patients. In group II drusenwere less in 7 out of 10 patients. In group III serous detachment disappeared in 3 out of 4 cases.Visual acuity was stable in all cases.

SRT MULTICENTER TRIAL

Because of the promising pilot study results, an international SRT multicenter trial has been start-ed to evaluate the therapeutic effect of SRT in patients with diabetic maculopathy and in pa-tients with macular edema after venous branch or central venous occlusion. Further few patientswith geographic atrophy secondary to AMD and patients with occult CNV secondary to AMD aretreated at single locations. Study centers are at the university eye clinics in Lübeck and Kiel andthe St. Thomas Hospital in London.So far, more than 60 patients with diabetic maculopathy were treated and controlled 6 monthsafter treatment. In 95 % of the patients the visual acuity was improved or stable at the 6 monthfollow-up control. The angiographic results with respect to leakage areas and Optical CoherenceTomography (OCT) data with respect to edema size and thickness however don’t show fully con-sistent with respect to the visual acuity.Further remarkable are the treatment of 10 patients with central serous rethinopathy (CSR) witha persisting edema for at least three month. For this pathology the results are very impressive.

Fig. 8. Fundus picture (A) and fluorescein angiogram (B) of an CRS patient. SRT lesions clearly show up in angiograp-hy.50 Test lesions for dosimetry were performed in the arcades, only highest pulse energies lead to a slightly visibleeffect. Treatment was performed in the macular area with pulse energies leading just to an angiographically visible ef-fect. For comparison to standard photocoagulation, (C) shows a fundus appearance after panretinal photocoagulation.21

62

All treated patients showed a complete disappearance of subretinal fluid after 4 weeks and 9 outof 10 patients demonstrated a significant increase in visual acuity after 3 months. These resultssuggest to change the standard recommendation for treating CSR patients. SRT offers the possi-bility of treating CSR and DMP patients much earlier in the course of the disease.

ON-LINE DOSIMETRY

The formation of microbubbles around the strong absorbing melanosomes inside the RPE hasbeen identified as the leading mechanism of RPE damage during µs laser pulse expo-sure.7,53 Mechanical effects such as cell disruption from a large number of simultaneously form-ing intracellular microbubbles most likely induce the RPE cell damage. If a correlation betweenfluorescein leakage and microbubble formation can be demonstrated, than the detection of mi-crobubbles enables an on-line dosimetry control avoiding test lesions and thus additional an-giography.If energy is absorbed and converted to heat, the thermoelastic expansion of the absorbing me-dium as a matter of principle leads to the emission of a bipolar pressure wave.54-55 Optoacoustic(OA) techniques have been used in ophthalmology for imaging of the ciliary body. OA monitoringof transscleral cyclophotocoagulation is possible in vitro.37 It can further be used to determinethe background RPE/retinal temperature increase during SRT.51 Using probe laser pulses it alsoenables for the first time to determine non-invasively on-line the temperature increase during cwphotocoagulation and during transpupillary thermo therapy (TTT), which could already be dem-onstrated in vitro. Further OA techniques are used for temperature mapping in tissue during la-ser induced thermo therapy (LITT).12-13, 28

Strong OA transients are also induced at the RPE during µs laser irradiation.51 After exceedingthe vaporization threshold, additional transients will be emitted owing to the formation and col-lapse of multiple microbubbles.49 Due to the µs heating and cooling and to the µs bubble life-time, both thermo-elastic and bubble induced pressure transients are expected in the ultrasonicMHz frequency range.By use of an ultrasonic needle hydrophone, we measured pressure in vitro during µs-laser expo-sure of the RPE.49,51 In order to non-invasively detect the acoustic transients during patient treat-ment a standard contact lens was modified with an ultrasonic transducer, as sketched in figure

Acoustictransducer

Acoustictransients Laser

pulses

Fig. 9. Contact lens with integrated ultrasonic pressure wave detector, after Schüle.49

63

9. With a transducer well adapted to the expected frequency range and amplitudes we were ableto detect pressure amplitudes on-line during SRT.Figure 10 displays typical superpositions of the 30 OA transients from each pulse of a pulsetrain as recorded with the contact lens, referring to two pulse energies and locations within onepatient’s eye. For exposures which result in no angiographic visible lesion (Fig. 10 A, 50 µJ) the30 OA transients show a clear reproducible bipolar thermoelastic pressure transient originatingfrom thermal expansion of the heated RPE. The pressure amplitudes are around 0.3 mbar. Athigher exposures (Fig. 10 B, 125 µJ, spot was angiographically visible) a stronger thermoelasticbipolar transient is measured. Moreover, small pulse-to-pulse fluctuations start after the first pres-sure peak most likely resulting from the statistic bubble formation and expansion. The delay tothe thermoelastic wave originates from the previous heating of the tissue and corresponds to theoptically detected bubble onset delay (Fig. 6). All treatment spots were ophthalmoscopically in-visible.A mathematical algorithm was derived to separate the fluctuations from the thermo-elastic back-ground.51 The algorithm gives a number which is related to the maximal pressure differencesbetween thermoelastic and bubble transients, the so-called OA-value. The analysis was limitedto a temporal window from 12 to 30 µs. Within this time frame most fluctuations appear.

ANGIOGRAPHIC RPE LEAKAGE VERSUS

OPTOACOUSTIC ON-LINE DOSIMETRY

Roider and Birngruber introduced angiographic determined fluorescein or ICG-leakage withoutophthalmoscopically visible effects as reference for selective RPE damage. In order to prove wheth-er the detection of microbubble formation correlates to angiographic findings, every single laserspot was analyzed with respect to its OA-value and angiographic response. The first study waslimited to four RCS patients, because due to the diffuse FLA leakage at the fundus in patientswith diabetic maculopathy a direct comparison of every single treatment spot was not possible.In figure 11 all analyzed OA-values of these four treatments are plotted over the pulse energyused. The open symbols mark all angiographically visible lesions, filled gray symbols the FLAnegative and filled black symbols the data points, which were not angiographically analyzable.A threshold for OA detectable angiographic lesion can be defined as OA = 1.96 10-10 bar × s.In case of these four treatments two out of 94 lesions were detected as false/positive and two asfalse/negative. The OA threshold for an angiographic lesion during patient treatment of

Fig. 10. Typical pressure transients from patient treatment: 30 pulses of one pulse train super-imposed, respectively.51

(A) Spot showing no RPE leakage, (B) Spot showing RPE leakage

64

OA = 1.96 10-10 bar × s is very close to the threshold of porcine RPE damage in vitro ofOA = 2.4 10-10 bar × s.51

By carefully conducting in vitro experiments it could be demonstrated that the extracted OA-valueincreased with the number of damaged cells within one spot.52 However, due to the limited opti-cal resolution, the amount of damaged RPE cells cannot be directly analyzed in a patient. There-fore, we plotted the OA-values measured for each spot and allocated them in figure 12. Roughlythe brightness of the spot increases with increasing OA-value, which might give a hint for anincreasing RPE damage. However, a quantitative analysis of the cell damage is impossible withcurrent techniques. Future investigations using ultra high-resolution OCT 17 in combination withadaptive optics might lead to the demanded resolution to quantify RPE damage and to followthe healing response in patients.In summary we proved that the concept of an OA-based non-invasive online dosimetry controlfor SRT is feasible. As well in vitro on porcine RPE samples as during SRT, the OA dosimetrysystem enables to detect the laser induced RPE damage. The OA dosimetry system is currentlyembedded in the multicenter clinical SRT study in order to control dosimetry, and is further eval-uated with respect to angiographic correlations.

ACKNOWLEDGEMENT

The authors like to thank Jörg Neumann and Dirk Theisen-Kunde from MLL, Georg Schüle (Lu-menis Inc, Santa Clara, CA), Carsten Framme (University Eye Clinic, Regensburg, Germany),Hanno Elsner (University Eye Clinic, Kiel, Germany), Peter Hamilton (St. Thomas Hospital, Lon-

Fig. 11. OA-value over applied pulse energy of four different CRS patient treatments. The open symbols mark the FLApositive lesions, gray symbols FLA negative and the black symbols not analyzable data points.51

65

don, Great Britain), Clemens Alt and Charles P. Lin (Wellman Center of Photomedicine and Har-vard Medical School, Boston, USA) for their participation in the various parts of the SRT project.

REFERENCES

(1) ANDERSON R.R., PARRISH J.A. − Selective Photothermolysis: Precise Microsurgery by Selective Ab-sorption of Pulsed Laser Radiation. Science 1983; 220: 524-527

(2) BIRNGRUBER R. − Thermal modeling in biological tissue. In: Hillenkamp, Pratesi, Sacchi (eds.) La-sers in Biology and Medicine. Plenum Press, New York 1980; 77-97

(3) BIRNGRUBER R., HILLENKAMP F., GABEL V.P. − Theoretical investigations of laser thermal retinalinjury. Health Phys. 1985; 48: 781-796

(4) BOULTON M.E., M X., KHALIQ A., MORIARTY P., J C., D M. − Changes in growth factor expressionin pig eyes following scatter laser photocoagulation. Invest Ophthalmol Vis Sci 1995; 36 (Suppl.): 95

(5) BRESNICK G.H. − Diabetic maculopathy: A critical review highlighting diffuse macular edema. Oph-thalmology 1983; 90: 1301-1317

(6) BRINKMANN R., HÜTTMANN G., RÖGENER J., ROIDER J., BIRNGRUBER R., LIN C.P. − Origin ofretinal pigment epithelium cell damage by pulsed laser irradiance in the nanosecond to microsecondtime regimen. Lasers Surg Med 2000; 27 (5): 451-464

(7) BRINKMANN R., KOOP N., ÖZDEMIR M., ALT C., SCHÜLE G., LIN C.P., BIRNGRUBER R. − Selec-tive damage of pigmented cells by means of a rapidly scanned cw laser beam. Proc SPIE 2002; 4617:134-140

(8) BRINKMANN R., KOOP N., ÖZDEMIR M., ALT C., SCHÜLE G., LIN C.P., BIRNGRUBER R. − Target-ing the Retinal Pigment Epithelium (RPE) by Means of a Rapidly Scanned Continuous Wave (CW)Laser Beam. Lasers Surg Med 2003; 32 (4): 252-264

(9) BÜLOW N. − The process of wound healing of the avascular outer layers of the retina. Light- andelectron microscopic studies on laser lesions of monkey eyes. Acta Ophthalmol (Copenh) 1978; 139(Suppl.): 7-60

(10) DOUKI T., LEE S., DOREY K., FLOTTE T., DEUTSCH T., DOUKAS A.G. − Stress-Wave-Induced Injuryto Retinal Pigment Epithelium Cells In Vitro. Lasers Surg Med 1996; 19: 249-259

(11) ELDIRINI A.A., OGDEN T.E., RYAN S.J. − Subretinal endophotocoagulation: A new model of subret-inal neovascularisation in the rabbit. Retina 1991, 11, 244-249

(12) ESENALIEV R.O., ORAEVSKY A.A., LARIN K.V., LARINA I.V., MOTAMEDI M. − Real-time optoacous-tic monitoring of temperature in tissues. Proc SPIE 1999; 3601: 268-275

Fig. 12. ICG-angiogram of different SRT spots. Numbers given are the determined OA-values (× 10-10 bar × s).62

66

(13) ESENALIEV R.O., LARINA I.V., LARIN K.V., MOTAMEDI M. − Real-time optoacoustic monitoring du-ring thermotherapy. Proc SPIE 2000; 3916: 302-310

(14) FIGUEROA M.S., REGUERAS A., BERTRAND J., APARICIO M.J., MANRIQUE M.G. − Laser photo-coagulation for macular soft drusen. Retina 1997; 17 (5): 378-384

(15) FRAMME C., SCHÜLE G., ROIDER J., KRACHT D., BIRNGRUBER R., BRINKMANN R. − Thresholddeterminations for selective RPE damage with repetitive microsecond laser systems in rabbits. Oph-thalmic surg las 2002; 33 (5): 400-409

(16) FRENNESSON C., NILSSON. − Significant decrease in exsudative complications after prophylacticlaser treatment of soft drusen maculopathy in a randomized study. Invest Ophthalmol Vis Sci 1997;38 (4 Suppl): 18

(17) FUJIMOTO J.G. − Optical coherence tomography for ultrahigh resolution in vivo imaging. Nat Biote-chnol 2003; 21 (11): 1361-1367

(18) GABEL V.P. − Die Lichtabsorption am Augenhintergrund. Habilitation Ludwig-Maximilians UniversitatMunchen 1974

(19) GABEL V.P., BIRNGRUBER R., HILLENKAMP F. − Visible and near infrared light absorption in pig-ment epithelium and choroid. Congress Series: XXIII Concilium Ophthalmologicum 1978; 450: 658-662

(20) GLASER B.M., CAMPOCHIARO P.A., DAVIS J.L., JERDAN J.A. − Retinal pigment epithelial cells re-lease inhibitors of neovascularization. Ophthalmology 1987; 94: 780-784

(21) GREHN F. − Augenheilkunde. Springer Verlag, Berlin - Heidelberg - New York 1998(22) HAGENAU R., BRINKMANN R. − Simulation of laser treatment on scalable parallel computer archi-

tectures. In: Jähnichen S., Zhou X. (eds.) The fourth international workshop on advanced parallelprocessing technologies. Ilmenau 2001; 137-146

(23) HERIOT W., MACHEMAR R. − Pigment epithelial repair. Graefe’s Arch Clin Exp Ophthalmol 1992;230: 91-100

(24) INOMATA H. − Wound healing after xenon arc photocoagulation in the rabbit retina. Ophthalmologica(Basel) 1975; 170: 462-474

(25) JOHNSON R.N., MCNAUGHT E.I., FOULDS W.S. − Effect of photocoagulation on the barrier functionof the pigment epithelium. II. A study by electron microscopy. Trans Ophthalmol Soc UK 1977; 97:640-651

(26) KELLY W.M., LIN C.P. − Microcavitation and cell injury in RPE cells following short-pulsed laser irra-diation. Proc SPIE 1997; 2975: 174-179

(27) KUCHLING H. − Taschenbuch der Physik. Fachbuchverlag, Leipzig 1996; 16 (Auflage)(28) LARIN K.V., LARINA I.V., MOTAMEDI M., ESENALIEV R.O. − Monitoring of temperature distribution

in tissues with optoacoustic technique in real time. SPIE 2000; 3916: 311-321(29) LATINA M.A., PARK C. − Selective Targeting of Trabecular Meshwork cells: In Vitro Studies of Pulsed

and CW Laser Interaction. Exp Eye Res 1995; 60: 359-372(30) LEWIS H., SCHACHAT A.P., HAIMAN M.H., HALLER J.A., QUINLAN P., VON FRICKEN M.A., FINE

S.L., MURPHY R.P. − Choroidal neovascularization after laser photocoagulation for diabetic macularedema. Ophthalmology 1990; 97: 503-511

(31) LIN C.P., KELLY M.W., SIBAYAN S.A.B., LATINA M.A., ANDERSON R.R. − Selective Cell Killing byMicroparticle Absorption of Pulsed Laser Radiation. IEEE J select Topics Quantum Electron 1999; 5(4): 963-968

(32) LORENZ B. − Morphologic changes of chorioretinal argon laser burns during the first hour post expo-sure. In: Welch (ed.) Lasers in the life sciences, Vol. 2. Harwood Acad Publ, Chur, London 1988;207-226

(33) MARSHALL J., CLOVER G., ROTHERY S. − Some new findings of retinal irradiation by krypton andargon lasers. In: Birngruber, Gabel (eds.) Laser treatment and photocoagulation of the eye. Junk Publ(Vol Doc Ophthalmol Proc Series), The Hague 1984; 36

(34) MATSUMOTO M., YOSHIMURA N., HONDA Y. − Increased production of transforming growth fac-tor-β2 from cultured human retinal pigment epithelial cells by photocoagulation. Invest OphthalmolVis Sci 1994; 35: 4645-4652

(35) NEUMANN J., BRINKMANN R. − Interferometric detection of laser-induced microbubbles in the reti-nal pigment epithelium. In: Birngruber R., Van den Bergh H. (eds.) Laser-Tissue Interactions, Thera-peutic Applications, and Photodynamic Therapy. Proc SPIE 2001; 4433: 81-86

(36) NEUMANN J., BRINKMANN R. − Microbubble dynamics around laser heated microparticles. ProcSPIE 2003; 5142: 82-87

67

(37) OBERHEIDE U., BRUDER I., WELLING H., ERTMER W., LUBATSCHOWSKI H. − Optoacoustic ima-ging for optimization of laser cyclophotocoagulation. J Biomed Opt 2003; 8 (2): 281-287

(38) PARRISH J.A., ANDERSON R.R., HARRIST T., PAUL B., MURPHY G.F. − Selective Thermal Effectswith Pulsed Irradiation from Lasers: From Organ to Organelle. J Invest Derm 1983; 80 (6): 75-80

(39) PLESSET M.S., PROSPERETTI A. − Bubble Dynamics and Cavitation. Ann Rev Fluid Mech 1977; 9:145-185

(40) PUSTOVALOV V.K. − Thermal processes under the action of laser radiation pulse on absorbing gra-nules in heterogeneous biotissues. Int J Heat Mass Trans 1993; 36: 391-399

(41) PUSTOVALOV V.K., KHORUNZHII I.A. − Theoretical consideration of the selective thermodenatura-tion of pigment epithelium layer using short laser pulses. Proc SPIE 1994; 2126: 111-115

(42) ROEGENER J., BRINKMANN R., LIN C.P. − Pump-probe detection of laser-induced microbubble for-mation in retinal pigment epithelium cells. J Biomedical Optics 2004; 9 (2): 367-371

(43) ROIDER J., MICHAUD N.A., FLOTTE T.J., BIRNGRUBER R. − Response of the Retinal Pigment Epithe-lium to Selective Photocoagulation. Arch Ophthalmol 1992; 110: 1786-1792

(44) ROIDER J., HILLENKAMP F., FLOTTE T.J., BIRNGRUBER R. − Microphotocoagulation: Selective ef-fects of repetitive short laser pulses. Proc Nat Acad Sci USA 1993; 90: 8463-8647

(45) ROIDER J., MICHAUD N., FLOTTE T., BIRNGRUBER R. − Histologie von Netzhautläsionen nach kon-tinuierlicher Bestrahlung und nach selektiver Mikrokoagulation des retinalen Pigmentepithels. Opht-halmologe 1993; 90: 274-278

(46) ROIDER J., BRINKMANN R., WIRBELAUER C., LAQUA H., BIRNGRUBER R. − Retinal sparing byselective retinal pigment epithelial photocoagulation. Arch Ophthalmol 1999; 117: 1028-1034

(47) ROIDER J., BRINKMANN R., WIRBELAUER C., LAQUA H., BIRNGRUBER R. − Subthreshold (retinalpigment epithelium) photocoagulation in macular diseases: a pilot study. Br J Ophthalmol 2000; 84:40-47

(48) ROIDER J., BRINKMANN R., BIRNGRUBER R. − Selective retinal pigment epithelium laser treat-ment. Theoretical and clinical aspects. In: Fankhauser, Kwasniewska (eds.) Lasers in ophthalmology:basic, diagnostic and surgical aspects. Kugler Publications, The Hague 2003; 119-129

(49) SCHÜLE G., HÜTTMANN G., ROIDER J., WIRBELAUER C., BIRNGRUBER R., BRINKMANN R. −Optoacoustic measurements during µs-irradiation of the retinal pigment epithelium. Proc SPIE 2000;3914: 230-236

(50) SCHÜLE G. − Mechanismen und On-line Dosimetrie bei selektiver RPE Therapie (dissertation, Medi-zinische Universität zu Lübeck) 2002

(51) SCHÜLE G., HÜTTMANN G., FRAMME C., ROIDER J., BRINKMANN R. − Non-invasive optoacoustictemperature determination at the fundus of the eye during laser irradiation. J Biomedical Optics 2004;9 (1): 173-179

(52) SCHÜLE G., ELSNER R., FRAMME C., ROIDER J., BIRNGRUBER R., BRINKMANN R. − Optoacous-tic real-time dosimetry for selective retina treatment. J Biomed Opt 2005; 10: 1-11

(53) SCHÜLE G., RUMOHR M., HÜTTMANN G., BRINKMANN R. − RPE damage thresholds and mecha-nisms for laser exposure in the µs to ms time regimen. Inv Ophthalmol Vis Sci 2005; 46 (2): 714-719

(54) SIGRIST M.W., KNEUBÜHL F.K. − Laser-generated stress waves in liquids. Journal of the AcousticalSociety of America 1978; 64 (6): 1652-1663

(55) SIGRIST M.W. − Laser generation of acoustic waves in liquids and gases. J Appl Phys 1986; 60 (7):85-121

(56) VOGEL A., BIRNGRUBER R. − Temperature profiles in human retina and choroid during laser coagu-lation with different wavelengths ranging from 514-810 nm. Lasers Light Ophthalmol 1992; 5: 9-16

(57) WALLOW I.H., TSO M.O.M., FINE B.S.F. − Retinal repair after experimental xenon arc photocoagu-lation. 1. A comparison between rhesus monkey and rabbit. Am J Ophthalmol 1973; 75: 32-52

(58) WALLOW I.H. − Repair of the pigment epithelial barrier following photocoagulation. Arch Ophthalmol1984; 102: 126-135

(59) WOLBARSHT M.L., LANDERS M.B. − The rationale of photocoagulation therapy for proliferative dia-betic retinopathy: A review and a model. Ophthalmic Surg Lasers 1980; 11: 235-245

(60) YAMAMOTO C., OGATA N., X Y., TAKAHASHI K., MIYASHIRO M., YAMADA H., UYAMA M., MATSU-ZAKI K. − Immunolocalization of transforming growth factor β during wound repair after laser photo-coagulation. Graefe’s Arch Clin Exp Ophthalmol 1998; 236: 41-46

68

(61) YOSHIMURA N., MATSUMOTO M., SHIMIZU H., MANDAI M., HATA Y., ISHIBASHI T. − Photocoagu-lated human retinal pigment epithelial cells produce an inhibitor of vascular endothelial cell prolife-ration. Invest Ophthalmol Vis Sci 1995; 36: 1686-1691

zzzzzz

corresponding address:

Dr. Ralf BrinkmannUniversity of Lübeck, Institute of Biomedical Optics andMedical Laser Center Lübeck GmbHPeter-Monnik-Weg 4, D-23562 Lübeck, GermanyTel: +49-451-500-6507brinkmann@mll.mu-luebeck.de

Prof. Dr. med. Johann RoiderUniversity of Kiel andClinic of Ophthalmology of the University Clinicum Schleswig-Holstein, Campus KielHegewischstr. 2, D-24105 Kiel, GermanyTel: +49-431-597-2360roider@ophthalmol.uni-kiel.de

Prof. Reginald BirngruberUniversity of Lübeck, Institute of Biomedical Optics andMedical Laser Center Lübeck GmbHPeter-Monnik-Weg 4, D-23562 Lübeck, GermanyTel: +49-451-500-6501bgb@mll.mu-luebeck.de

69